Wednesday, 2017-01-04

*** Syed__ has quit IRC01:25
*** VW has joined #craton02:26
*** VW has quit IRC02:30
*** rainya has joined #craton02:31
*** rainya has quit IRC02:31
*** tojuvone has quit IRC04:08
*** VW has joined #craton05:00
*** VW has quit IRC05:55
*** jovon has joined #craton06:14
*** tojuvone has joined #craton06:39
*** VW has joined #craton07:55
*** VW has quit IRC08:00
*** jovon has quit IRC08:40
*** VW has joined #craton14:45
*** VW has quit IRC14:48
*** VW has joined #craton14:48
*** Mudpuppy_ has joined #craton14:58
*** VW has quit IRC15:08
*** VW has joined #craton15:09
*** VW has quit IRC15:13
*** Mudpuppy_ has quit IRC15:13
*** VW has joined #craton15:13
*** Mudpuppy_ has joined #craton15:13
*** valw has joined #craton15:21
*** valw_ has joined #craton15:27
*** valw has quit IRC15:28
*** mudpuppy has quit IRC15:28
*** mudpuppy has joined #craton15:29
*** Syed__ has joined #craton15:53
*** valw_ has quit IRC16:08
*** valw has joined #craton16:09
*** valw has quit IRC16:30
*** valw has joined #craton16:36
*** valw has quit IRC16:59
*** VW has quit IRC17:10
*** VW has joined #craton17:11
*** VW has quit IRC17:16
*** VW has joined #craton17:22
*** VW_ has joined #craton17:48
*** VW_ has quit IRC17:49
*** VW_ has joined #craton17:49
*** VW has quit IRC17:51
*** VW_ has quit IRC18:20
*** VW has joined #craton18:22
*** VW has quit IRC18:22
*** VW has joined #craton18:23
*** valw has joined #craton18:33
sigmavirusjimbaker: ping18:50
*** hughsaunders has quit IRC19:36
*** d34dh0r53 has quit IRC19:36
*** hughsaunders has joined #craton19:36
*** d34dh0r53 has joined #craton19:36
*** valw has quit IRC19:36
*** valw has joined #craton19:37
jimbakersigmavirus, hey19:39
sigmavirusHave you looked at the spec that's up?19:45
*** VW_ has joined #craton19:51
*** VW_ has quit IRC19:52
*** VW_ has joined #craton19:52
*** VW has quit IRC19:54
jimbakersigmavirus, i have19:54
*** VW_ has quit IRC19:56
jimbakersigmavirus, so page/per_page or the similar limit/offset seems to correspond best to what sqlalchemy provides; i understand the motivation behind limit/marker, but i don't know how one implements the desired stability in general against our APIs without maintaining a cursor20:00
jimbakerso i'm certainly curious about that aspect20:01
*** valw has quit IRC20:05
*** valw has joined #craton20:06
*** VW has joined #craton20:07
*** valw has quit IRC20:09
jimbakersigmavirus, here's one option, a project to help support limit/marker - https://github.com/djrobstep/sqlakeyset20:17
jimbakerchecking if there's anything similar in oslo db20:17
jimbakeryes there is this: http://docs.openstack.org/developer/oslo.db/api/oslo_db.sqlalchemy.utils.html#oslo_db.sqlalchemy.utils.paginate_query20:18
Syed__jimbaker: do you have some time to talk about oslo20:18
jimbakersigmavirus, so my quick read here it is readily doable by working with arbitrary queries. nice.20:19
jimbakerSyed__, you mean olso.policy?20:19
Syed__yeap for rbac support20:19
sigmavirusjimbaker: yeah, looks that way. Was researching it as we spoke20:19
* sigmavirus knew it had to be in there somewhere20:19
jimbakergot it. yeah, let me first get lunch20:19
Syed__sure20:20
Syed__thanks20:20
jimbakerso maybe we can discuss in about an hour. 2:30p MT?20:20
Syed__Sounds good i.e. 3:30p SA time..20:21
Syed__sorry i really need to get the hang of mountain times, central times etc.20:21
jimbakersigmavirus, cool. yeah i think that util function looks generally useful - we always have some guiding model we care about, vs some arbitrary vie20:21
jimbakerSyed__, yes, that's why we generally use UTC instead for open source ;)20:21
jimbakerthen it complicates things for both parties, unless they live in iceland :_)20:22
jimbaker*arbitrary view*20:23
jimbakerok, biab20:23
sigmavirusSyed__: yeah, we're 1hr ahead of mountatin time20:26
*** VW has quit IRC20:26
sigmavirusmountain time too ;)20:26
*** VW has joined #craton20:27
*** valw has joined #craton20:47
jimbakerSyed__, can we push this meeting out by an hour? my car was just fixed, and i want to get it home given the snow storm we are having21:27
*** VW has quit IRC21:33
*** VW has joined #craton22:03
*** VW has quit IRC22:06
*** VW has joined #craton22:06
*** Mudpuppy_ has quit IRC22:17
Syed__sure, let me know whenever you are back jimbaker22:22
*** VW has quit IRC22:40
*** valw has quit IRC22:42
*** VW has joined #craton22:44
*** VW has quit IRC22:51
*** VW has joined #craton22:51
*** openstack has joined #craton22:54
*** VW has quit IRC22:57
*** VW_ has quit IRC22:58
jimbakerSyed__, back23:01
jimbakerSyed__, i love snow but it certainly delays things :)23:01
Syed__haha, i hope you got your car back safe23:02
jimbakerjust fine23:02
Syed__jimbaker: yeap, so was going through http://docs.openstack.org/developer/oslo.policy/usage.html#migrating-to-oslo-policy23:03
Syed__and making a policy file for oslo policy to be used for craton23:03
jimbakerok23:03
Syed__using this https://github.com/openstack/nova/blob/master/nova/policy.py as a reference23:03
Syed__should we use nova to authenticate or how the authentication should be working is my question23:04
jimbakerSyed__, you mean keystone23:05
Syed__and we also do need http://docs.openstack.org/kilo/config-reference/content/policy-json-file.html23:05
Syed__oh sorry yeah keystone, so keystone is being used as authentication23:05
jimbakerSyed__, we can either use keystone or craton's authentication23:06
Syed__Shouldn't we be using just keystone for this purpose ?23:06
jimbakerno, otherwise we don't have rbac for the scenario without keystone23:06
jimbakerthat would be unacceptable imhop23:06
jimbakerimho23:06
Syed__i see .. Okay so we need to make sure  that authentication works with craton and keystone23:07
jimbakeryeah, this should be just fine23:07
jimbakerthere's nothing in oslo.policy that depends directly on keystone23:07
Syed__okay, and ones we have basic layout we can add user scenarios to test cases as they are in that blueprint23:07
jimbakerSyed__, i would look at the unit tests for oslo.policy in projects like cinder23:07
Syed__Okay, yes i will go look that out right now23:08
jimbakerbest bet: try stuff out in the python console23:08
Syed__that is what i have been doing lately23:08
jimbakeryou don't want anything that has more than a couple of imports23:08
Syed__but thanks, will need to dig into how unit testing is done in other projects for oslo, will be doing that rn23:08
jimbakerwant to make the scenario as simple as possible :)23:09
Syed__cool, thanks23:09
jimbakerSyed__, i'm going to look at this with you a bit more today... one moment23:11
Syed__should we hop into one of the vidyo rooms ?23:12
*** VW has joined #craton23:14
jimbakerSyed__, nahh, i think irc is much better here23:18
*** VW has quit IRC23:18
jimbakervideo is overrated :)23:18
jimbakermuch easier to discuss text when using text23:19
Syed__i am looking at this https://github.com/openstack/cinder/blob/4962723c9a1da7d2a75e2b63eccb63619dac22b6/cinder/tests/unit/conf_fixture.py23:20
jimbakerSyed__, yeah, i'm looking for a better example23:21
jimbakeri should have made a better note of which test i found gave what i think is the best approach23:21
jimbakeri'm sure i will re-find momentarily!23:22
jimbakerSyed__, i guess it's in keystone after all. so take a look at https://github.com/openstack/keystone/blob/master/keystone/tests/unit/test_policy.py#L6123:27
jimbakerdespite being in keystone23:27
jimbaker's test suite23:27
jimbakerit does not rely on any keystone setup23:27
jimbaker(after all, unit tests!)23:27
Syed__looking23:28
jimbakerso some examples of rules, as set in self.rules23:28
jimbakerso may not be the best testing, because not comprehensive. it's possible that oslo.policy's own tests are better, but they are also fairly abstract about what is going on23:30
jimbakerbut take a look at stuff like https://github.com/openstack/keystone/blob/master/keystone/tests/unit/test_policy.py#L11823:31
Syed__hmm23:31
Syed__i looked into magnum policies as well, but i guess this is more helpful with what we want to be done23:34
Syed__will be looking at this more23:34
jimbakerSyed__, so what you want to see here is: i have some credentials23:34
jimbakerand we forward chain to the desired target - or not23:38
jimbakeri guess the usual usage is http://docs.openstack.org/developer/oslo.policy/api/oslo_policy.html#oslo_policy.policy.Enforcer.authorize, but the tests here use enforce23:38
jimbakerprobably no diff for unit testing23:38
jimbakerSyed__, i'm putting together a simple script based on what's in that keystone test23:49
jimbakerjust so we can try things out without the distraction of keystone specifics23:50
jimbakerSyed__, one fun fact ;) is that the keystone.policy.backends.rules.enforce function switches the order of args compared to the method it wraps, oslo_policy.policy.Enforcer.enforce23:57
jimbakerjust because it can...23:57
jimbakerno doubt keystone did it one way, and it just stuck23:58
jimbakeranyway, that's part of the simplification necessary here23:58

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!